چكيده به لاتين
Abstract:
In this thesis, a robust estimator has been proposed to estimate the vehicle longitudinal velocity. Longitudinal velocity has an important role in revealing slip and increase safety of transportation system. Since the conventional methods of longitudinal velocity sensors such as accelerometers, in addition to being costly, have low reliability, hence there are a lot of research in the case of estimating vehicle longitudinal velocity for calculation of slip value and use it in vehicle stability control and deduction of braking distance. In the most of these researches, model of system has an important role in the accuracy of estimator performance and since finding a model with high accuracy and ability to consider uncertainties and turbulence effect is difficult, this research proposes a distributed estimator via multiple models which is close to real dynamic model of vehicle and has high accuracy. In this proposed multiple models, the complex and nonlinear model of system will be divided to some linear model in different working condition of vehicle. So several high accuracy linear model will be used to estimate vehicle longitudinal velocity. Because of existing matrix constraints in the estimating part, we use Linear Matrix Inequality (LMI) for estimator stability. After estimating longitudinal velocity of one wheel, this procedure will be used for both wheels and a unique and accurate estimate will be obtained via data fusion algorithm. Simulation results will be proposed in order to present accuracy of new estimator and the validity of this research has been confirmed after experimental test on an vehicle. We use the model of automobile because we have the real wheel speed and longitudinal velocity through experimental test.
Keywords: Vehicle Velocity Estimation, Multiple Models theory, Wheel Slip, Estimator Stability
